Investigation of fall risk in patients with Parkinson's disease and establishment and validation of a nomogram prediction model based on LASSO regression
10.13406/j.cnki.cyxb.003926
- VernacularTitle:帕金森病患者跌倒风险调查及基于LASSO回归的列线图预测模型建立与验证
- Author:
Haiqiong HU
1
;
Lixia LI
;
Yu SHAO
;
Yuanyuan HUANG
;
Fajun XIAO
;
Ke XIA
Author Information
1. 四川省南充精神卫生中心神经内科,南充 637000
- Keywords:
Parkinson's disease;
fall;
multivariable analysis;
nomogram model
- From:
Journal of Chongqing Medical University
2025;50(10):1338-1344
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To investigate the risk of falls in patients with Parkinson's disease and establish and validate a prediction model.Methods:We selected 372 patients with Parkinson's disease at Sichuan Nanchong Mental Health Centre from January 2022 to Septem-ber 2023.The patients were divided in a 7:3 ratio into model group(260 cases)and validation group(112 cases).According to previ-ous literature and suspected factors found in clinical practice,we collected general information(sex,age,etc.)and disease-related factors(the duration of Parkinson's disease,the type of medications taken,etc.)that may be associated with falls in patients with Parkinson's disease.In the model group,between patients with and those without falls within 1 year as reported by the patients or their family members,potential predictors for falls were determined through comparison of general information and disease-related factors,least absolute shrinkage and selection operator(LASSO)regression,and multivariable logistic regression.Based on the significant factors,a nomogram model was established and validated.Results:In the model group,81(31.15%)of the 260 patients experienced falls.According to the LASSO regression and multivariable logistic regression results,alcohol consumption,the type of medications,the score of the Unified Parkinson's Disease Rating Scale part Ⅲ(UPDRS-Ⅲ),the Berg Balance Scale score,the presence of arthritis,and the presence of osteoporosis were independent factors influenc-ing falls in patients with Parkinson's disease.The area under curve(AUC)of the receiver operator characteristic curve(ROC)for pre-dicting the risk of falls in patients with Parkinson's disease was 0.896(95%CI=0.856-0.935)in the model group and 0.883(95%CI=0.840-0.926)in the validation group.The calibration curve analysis results showed that the prediction curves of the model and valida-tion groups closely fitted the standard curves.The decision curve analysis results indicated that when the probability threshold for pre-dicting the fall risk in Parkinson's disease using the nomogram was 0.10-0.90,the net benefit rate of the patients was greater than 0.Conclusion:The risk of falls in patients with Parkinson's disease are mainly influenced by factors such as alcohol consumption and the type of medications.The nomogram model established in this study can be used to predict the fall risk in patients with Parkinson's disease.
